JSON-LD Schema for AI Visibility

JSON-LD schema markup increases AI citation rates by 19 percentage points

Pages with JSON-LD structured data achieve a 47% Top-3 citation rate in Perplexity compared to 28% for pages without it — a 19-percentage-point advantage. Structured data tells AI retrieval systems exactly what your page is about, who wrote it, and what questions it answers, making it significantly easier for AI to select your content as a citation source.

What is JSON-LD?

JSON-LD (JavaScript Object Notation for Linked Data) is a method of embedding machine-readable metadata directly into your web pages using Schema.org vocabulary. Unlike microdata or RDFa, JSON-LD sits in a script tag in your page's head — it doesn't interfere with your HTML structure and is invisible to users.

Search engines and AI crawlers read JSON-LD to understand the semantic meaning of your content: what type of page it is, who created it, what organisation it belongs to, and what questions it addresses.

The five essential schema types for AI visibility

1. Organization schema

Defines your business entity — name, address, logo, contact information, social profiles, and area served. This is the foundation for entity recognition. Without Organization schema, AI tools may struggle to associate your content with your brand.

2. Article schema

Applied to blog posts and knowledge articles. Includes author information, publication date, modification date, headline, description, and publisher. AI retrieval systems use dateModified to assess content freshness — a critical ranking signal where 70% of top citations come from content updated within 18 months.

3. FAQPage schema

Structures question-and-answer pairs in a format AI models can directly extract. FAQ schema content is particularly well-suited for AI citation because the question-answer format matches how users query AI tools. Google also displays FAQ schema as rich results in search.

4. HowTo schema

Step-by-step instructions with named steps, tools required, and estimated time. Process-oriented content with HowTo schema maps directly to how-to queries in AI search, making it more likely to be retrieved and cited for instructional queries.

5. LocalBusiness schema

For businesses serving a geographic area. Includes address, opening hours, service area, price range, and reviews. Critical for location-based AI queries like "best accountant in Parramatta" or "cybersecurity company Sydney".

How AI models use structured data

AI retrieval systems process JSON-LD at three stages:

1. Indexing — schema helps the crawler categorise and embed your content correctly. A page with Article schema is indexed as an article, not a product page or generic web page
2. Retrieval — when a user query matches the entities and topics defined in your schema, your page is more likely to be retrieved as a candidate source
3. Reranking — schema provides trust signals (author expertise, publication date, organisation credentials) that improve your score during the quality reranking phase

Implementation checklist

• Every page — Organization schema in the site layout (global)
• Blog articles — Article schema with author, datePublished, dateModified
• Service pages — Service schema with provider, serviceType, areaServed
• FAQ sections — FAQPage schema wrapping each Q&A pair
• Contact page — LocalBusiness schema with full NAP details
• How-to guides — HowTo schema with named steps

Common mistakes that reduce effectiveness

• Missing dateModified — without this, AI models can't assess freshness. Always include and keep it current
• Generic author — "Admin" or "Staff Writer" provides no expertise signal. Use real names with credentials
• Schema without matching content — if your FAQPage schema lists questions that aren't answered in the visible content, search engines may flag this as spam
• Duplicate Organization schema — use one consistent Organization schema across all pages, not different versions on different pages
• Missing sameAs links — your Organization schema should include sameAs links to all official profiles (LinkedIn, YouTube, industry directories) to strengthen entity recognition

Testing your schema

After implementing JSON-LD, validate it using:

• Google Rich Results Test — search.google.com/test/rich-results
• Schema.org Validator — validator.schema.org
• Browser DevTools — search for application/ld+json in page source to verify rendering